DM32 Curve Fit

CFIT Program from the Owner's Manual

The listing from the Owner's Manual has been entered into the calculator and saved as a State File

cfit.d32

The program provides the usual set of Curve Fitting for Straight Line, Exponential, Logarithmic and Power curves, automating the same techniques as the manual method for DM15 described here DM15L Non-linear Curve Fit .

There is no “best fit” algorithm, which means you need to decide in advance which of the 4 curve types to fit, before entering the data. If you choose a different curve you have to enter the data again.

Each curve type has a separate program label

• XEQ S : Straight Line
• XEQ E : Exponential
• XEQ L : Logarithmic
• XEQ P : Power

Data is entered the same way whichever of the 4 programs is chosen.

• X-value R/S Y-value R/S
• X-value R/S Y-value R/S
• X-value R/S Y-value R/S
• X-value R/S Y-value R/S

Once all the data is entered you run the fit using XEQ R

This gives the various paramaters

• R-value (correlation coefficient) R/S
• B-value R/S
• M-value R/S
• Prompt for an x-value to use to calculate ŷ - enter a number R/S
  • ŷ is shown for that x-value
• Prompt now for a y-value to use to calculate an estimate of x R/S
  • x estimate is shown for that y-value
• Prompt for a new x-value to use to calculate ŷ R/S
  • this goes on until you want to stop….
  • press C to quit

• Straight Line y = B + Mx
• Exponential y = BeMx
• Logarithmic y = B + M Ln(x)
• Power y = BxM

• XEQ L
• 0.3 R/S 1 +/- R/S
• 1 R/S 2 R/S
• 2 R/S 3.8 R/S
• 3 R/S 5 R/S
• 4 R/S 5.6 R/S
• 5 R/S 6 R/S

• XEQ R
• R=0.99932 R/S
• B=2.05210 R/S
• M=2.5359 R/S
• enter an x-value to calculate an estimate of y
  • 2.5 R/S
  • y? 4.37571
• enter a y-value to calculate an estimate of x
  • 0 R/S (what is x-value where the curve crosses the x-axis at y = 0? The solution to the equation)
  • X ? 0.44520
• finish C

• The curve equation is y = 2.05210 + 2.5359Ln(x)
• The values for R, B and M are stored in register-variables R, B and M for future use
  • e.g. make an EQN
  • Y = B + M x Ln(X)
  • the equation can be evaluated, solved or integrated using the stored values for B and M
• Until a new Curve fit program is run, or anything is done to change the Statistics Registers, you can re-run the fit with XEQ R to allow more y and x estimates to be made

Page created : 17/04/26 08:35 BST

Page updated : 01/01/70 01:00 BST